Linking the human nervous system and brain directly to a computer opens up innumerable possibilities, not only in the future world of medicine, but also as a potential way of technically evolving all humans. This, however, presents something of an ethical problem. Nevertheless, the only way to actually find out what is realistically possible and what is not is to carry out practical experimentation using implant technology and to witness the results.
Ken Warwick (Warwick 2005)
Implants have a general disadvantage over other methods employed to restore or even enhance neural function: they involve a more or less invasive surgical procedure. Still, their development has been pursued with considerable perseverance for a number of important reasons, which we have to keep in mind when wondering why anybody would seriously consider surgical connection of artificial devices to the human brain: Neural implants
- can restore neural function where all other methods fail (e.g. CI, ABI, retinal implant),
- can function continuously without the implantee having to pay any attention to them or having to interrupt or alter his or her normal behaviour,
- can be completely hidden under the skin, invisible to others,
- can be turned off easily,
- can be recharged without having to remove the implant.
To judge the actual weight of these advantages, we ought to take a closer look at the current limitations in the field of central neural implants. While auditory implants have been improved considerable over the last 40 years, and visual implants appear on the clinical horizon, there are still no prosthe-ses for the restoration of the sense of taste, smell and touch yet. The latter has already been addressed (Sampaio et al. 2001, Tyler et al. 2003; Krupa et al.
2004, Rita 2004) and - in combination with motor prosthesis - the bionic arm is no longer pure fiction as encountered in Schwarzenegger movies.
The first limitation affecting all sensory implants today is that they contact the neural tissue with a relatively small number of electrodes as compared to the multitude of neurons involved in the sensory pathways. Larger electrode arrays are currently under investigation for implantation in the visual cortex (Normann et al. 2001; Fernandez et al. 2005). However, the more electrodes become available on a small space, the more cross channel interaction is to be expected, possibly leading to a deterioration of the performance of the implant.
Second, implants today are placed in sensory pathways that have been severed before. With a lesion in the central nervous system there usually is little chance of natural regeneration. To the contrary, other elements in the severed pathway degenerate, too, when not in use. An early station in the auditory pathway for instance, the spiral ganglion, may lose as many as 75% of its cells when the hearing nerve is severed. The same is true for the peripheral sensory and motor elements one a peripheral nerve is cut off. Degeneration may be delayed by delivering protective substances or continuous electrical stimulation to the structures at risk of degeneration. There are also attempts to place implants at a higher level in the sensory pathway (e.g. the colliculus inferior or the hearing cortex in the auditory path) where degeneration after nerve injury is minimal.
Third, the electrodes contacting the neural tissue are prone to rejection and degradation. They may also damage the neural tissue they are supposed to stimulate. Therefore, researchers are looking into ways of coating of the materials in vitro, either with epithelial cell layers or with non-degradable surface layers.
Fourth, the neural interfacing of electrodes is still far from mimicking the anatomical and physiological connections in a neural network. To improve the performance of auditory brainstem implants, for instance, it may be of advantage to place the electrodes as close to their neural target structures.
This is the principle behind current research with penetrating brainstem electrodes for the cochlear nucleus (Schindler 1999; McCreery et al. 2000; Rosahl et al. 2001; Rosahl 2004).
From the perspective of today's technology, however, it is hard to see how the electrochemical basis of thousands of synapses can be replaced by an electronic system. It might not be reasonable at this point even to try to "rebuild", completely, a synaptic network. To restore sensory pathways, "bioinspired" systems in hearing and visual prostheses (Fernandez 2000; Fernandez 2002, Fernandez et al. 2005) partially mimic the function of biological receptors and sensor cells by "translating" signal received from the environment into physiologically "acceptable" electrical stimulation patterns that can be relayed to neurons in the remaining intact part of sensory pathway information via electrode arrays. At that point the intact portion of the central nervous system takes over to process this information further in its usual way. To improve the performance of a sensory implant, one therefore has to improve the way the information is delivered to the first intact neurons. Ideally, the implant would have to excite and inhibit these neurons in a pattern identical to the physiological paragon. This would involve several thousand electrode contacts in a volume of a few cubic millimeters - a setup that is unlikely ever to be realised. Alternatively, the (electrical) input/output functions of a group of neurons may be simulated, so that an array of electrodes connected to neuronal tissue on one side and a signal processor on the other would just have to deliver stimuli in a charge pattern that equals the one that is present in a given group of neurons at a given time in response to a given environmental stimulus. This is, basically, where research and clinical application stands today. It remains to be seen to what extent it will be possible with these systems to completely restore normal sensory pathways by neuroelectronic interfaces. However, the better these interfaces become, the more one could imagine that electronic connections to the brain that allow contacting biological structures involved in memory formation and higher cortical functions can be established one day - be it merely by modulating the amount of neurotransmitters in a specific area of the cerebral cortex or in a functionally specific neuronal network. Improvement of these functions would be particularly welcome for patients with neurodegenerative disorders like Alzheimer's disease - but if there were ever a very low-risk way to extend human memory capacity by directly connecting it to artificial devices with a, it is hard to see how healthy people too would not want to take advantage of such a possibility (see section on possible solutions below).
As of today, technology has not advanced that far and there are no such devices on the horizon yet. In a functional retina implant, for instance, colour vision would not be possible since all visual information is converted into a grey scale. Also, the understanding of speech and the perception of music with any multi-channel implant of today's technological generation connecting to a secondary or tertiary neuron of the auditory pathway will probably remain rudimentary, as has already been demonstrated in auditory brainstem implants.
Stimulation that is purely based on electrical impulses is further limited by the refractory properties of neurons, which respond only to a limited number of impulses per time interval. Perhaps one of the most important limitations in research on sensory implants is, however, that their efficacy can ultimately only be tested in humans. On the other hand, it is exactly this consideration that also poses a temptation that researchers could not always resist in the past. As soon as commercial interests will be involved, ethical dams may well be in danger of breaking (see discussion on ethical issues below). Such interests may arise in particular with the development of devices that could not only restore severed sensory function, but computer-enhance human capabilities.
For Human Machine Interfaces, particularly for the control of artificial limbs, crucial limiting factors are the spatial and temporal resolution of the electrical brain activity that reflects the ongoing neuronal activity, as well as the speed and accuracy of three-dimensional movements to be performed in response to these brain signals. Also, sensory (haptic) feedback from prosthetic arms can not be provided yet, and motor functions are restricted to very basic movements so far.
Size, biocompatibility, durability, and energy supply are basic problems for all neuroelectronic implants, but considering recent developments, these do not appear to remain critical in the long run. It is still not possible to connect more complex brain structures such as the hippocampus - a formation of neural tissue in the temporal lobe involved in memory formation and retrieval - or speech centres with electronic circuitry, although research studies in that direction are currently under way (Iatrou et al. 1999; Alataris et al. 2000; Gholmieh et al. 2002; Gholmieh et al. 2004) and will be dealt with later in this section.
Scientists have adopted several strategies to overcome these present limitations. Some of such strategies will be summarised briefly in the next section to provide an understanding of current top-of-the-line research in this field.
Possible solutions. Speed as a limiting factor in the performance of BCIs can be increased even without manipulating the interface itself, just by "tuning" the output functions on the technical part. For instance, with a pointing device or a virtual keyboard connected to a BCI, it may be possible to predict an intended target from an early anticipated trajectory and to place the cursor or pointer at the predicted endpoint. Recent research has indicated that with such a higher performance 96-electrode BCI it is possible to design a fast and accurate key selection system with a speed of 6.5 bits per second or approximately 15 words per minute(Santhanam et al. 2006).
When considering improvement of the quality of complete neural-electronic interfaces, researchers currently employ two different approaches. The first approach is to obtain more information on the structural organisation and the working principles of neuronal networks and their function, as has recently been demonstrated for the mechanisms of grasping and movement intention (Carpenter 2002; Cattaneo et al. 2005; Fogassi and Luppino 2005). Another example in that line has inspired much enthusiasm in the scientific community. The experiment, conducted by researchers at the University of California in Berkeley, involved neuronal signals from 117 neurons in a relay station of the visual pathway close to the thalamus (lateral geniculate ganglion, LGL) in a cat. The LGL is connected directly to the cat's eyes via the optic nerve. Each of its cells is programmed to respond to certain features in the cat's field of view. Some cells "fire" when they record an edge in the cat's vision, others when they see lines or different patterns at certain angles. Using a technique the researchers describe as a "linear decoding" to recon struct an image from their data, they saw natural scenes with recognisable moving objects (Stanley et al. 1999). Research at the same laboratory also demonstrated that it is possible to map non-linear neuronal responses to visual stimuli in the visual cortex (Lau et al. 2002). Knowledge on these response characteristics can be applied to identify the major input/output pathways of a specific functional pathway in the brain.
Frank Ohl of the Leibniz Institute for Neurobiology has extensively studied the functional organisation of the auditory cortex in gerbils. His research group recorded electrical brain activity in response to rising and falling "tones". When the auditory cortex itself was stimulated with a similar electrical signal, the animals were able to discriminate rising and falling tones created by direct electrical stimulation in the absence of any sound presentation (Ohl et al. 2001).
The second approach actually builds upon the former by attempting to model neurobiological structures that have already been extensively studied before ("morphing"). One example is the recent development of an artificial retina. "Visiol", a silicon-based microchip that includes 3600 output "cells" simulating the response characteristics to light stimulation of the four major clusters of retinal ganglion cells, was designed and built at the University of Pennsylvania in 2001. With their axonal processes, these cells account for 90% of the fibres of the optic nerve (Zaghloul and Boahen 2004a; Zaghloul and Boahen 2004b). The "neuromorphic" chip needs only 60mW electrical power - one thousandth of the power required by a regular PC. This chip may pave the way to a complete ocular prosthesis, including camera, processor, and stimulator.
In a similar way, "neural morphing" may be employed to simulate the input/output-functions of other neuronal networks, such as the auditory system, without the need to adhere strictly to the biological principles realised within a neuronal network. It has been demonstrated that the tone pitch gradients of the cochlear nucleus - an assembly of neurons in the brain stem - can be accessed through electrical stimulation with microelectrodes on the brain's surface (McCreery et al. 1998; McCreery et al. 2000). Similarly, more complex brain circuits and functions may be accessible for establishing a "dialogue" between a computer and the brain in the future (Ohl et al. 2000, Ohl et al. 2003a).
One of the most ambitious projects in this respect combines methods of computational neuroscience with computer engineering to "morph" functions of the hippocampus with computer hardware and software. Theodore Berger, a professor of biomedical engineering at the University of Southern California, and his team succeeded in simulating some basic non-linear functions of this complex structure with computer equipment (Gholmieh et al. 2001; Chauvet and Berger 2002; Gholmieh et al. 2004; Berger et al. 2005). Their ultimate goal is an "artificial hippocampus" that could restore or even enhance memory.
Unlike devices such as cochlear implants, which merely stimulate brain activity, this chip implant would have to perform the same processes as the damaged part of the brain it is replacing. It would have to receive input from the brain and to deliver its output to the brain. At present, it is hard to see how such an interface could be established. Also, human memory is not static. It is based on continuous changes in the efficacy and qualities of cellular and molecular processes, so the implant device may need to adapt to such changes. Considering the complexity of the task, it will probably be a long way from Berger's initial results to the actual re-establishment of lost memory function, not to mention enhancement of memory in perfectly healthy people.
In the meantime, clinical applications of electrical brain stimulation have gained world-wide acceptance in the treatment of movement disorders. Indications have been extended and Medtronic Inc., the major manufacturer of such implants, uses its products in clinical tests to pulse the thalamus for the treatment of epilepsy; another region of the deep brain to treat migraines, depression, and obsessive-compulsive disorder; the hypoglossal nerve in the neck to treat sleep apnea; the sacral nerve to treat bowel disorders; and the stomach to treat obesity. Parallel to the efforts of modelling and duplicating neural networks, efforts are underway to establish improved connections between electrodes and neurons.
Microtechnology now allows the electrodes to be miniaturised, and their active surface can be enlarged by laser treatment. Coating of the electrodes with growth factors appears to improve the electrical contact to the neural structures, and conductive tarnish with nanoparticles reduces the breakdown of electrodes by living tissue. Neurons cultured on nanofibres develop neurite extensions and the artificial material counteract astrocytic scar formation, at the same time (McKenzie et al. 2004; Webster 2004). Also, neural cells can be grown onto silicon chips, contact them, and promote fibre growth connecting the implant to the nervous system.
Neural implants capable of communicating with the nervous system on a chemical rather than electrical basis (electro-osmosis) are being developed at the University of Stanford. Such "microfluidic" chips can serve both as neural prosthesis (artificial retina) and focal drug delivery systems. The way they are set up is a combination of electronic implants as described earlier, with more or less mechanical actuators capable of releasing a variety of different chemicals, e.g. neurotransmitters or drugs, in a very small volume of biological tissue. Medication could be delivered topically through micropumps implanted over the cerebral cortex where it is released in response to intracortically detected changes in the electroencephalogram (EEG) that may predict an incipient seizure in epileptic patients.
To prevent a generalised seizure it may also be possible to stimulate electrically a cortical brain area upon EEG changes - in fact, a system with eight electrodes connecting to "hot spots" of epileptic activity has already been implanted at Rush University in Chicago in 2004 (NeuroPace Inc., Mountain View, California, USA) (Worrell et al. 2005). For these purposes, passive implants will be replaced by active ones that have microprocessors connected to the implanted unit or even directly adjacent to the electrodes.
Although implants resulting in enhancement are not part of today's clinical reality, they are a realistic future option for which we should be ethically prepared.
While current technology with neural implants based on electrode (or electro-osmotic) - tissue interfaces may be of help for patients to overcome a severe neurological impairment, their potential to damage normal tissue has so far prevented application in healthy persons, with a few exceptions involving human self-experimentation (Warwick 2002; Warwick et al. 2003; Warwick 2005).
However, as efficient electrical and microfluidic implants become available, the next step after therapy may be enhancement. The definition of the term "enhancement" is not unequivocal. Hall refers to neuroenhancement as "the use of drugs or other interventions to modify brain processes with the aim of enhancing memory, mood and attention in people who are not impaired by illness or disorder" (Hall 2004). This definition, however, does not cover sensory and motor enhancement with neural electronics. Moreover, neural implants (as opposed to drugs) would not necessarily have to "modify brain processes" to improve human performance. They would rather build on the normal function, and access input and output, of per se unmodified neuronal networks. In this context, a "modification" of brain function would rather appear to be a side effect of such an implant.
The Council on Bioethics to the U.S. President offers a different definition by stating that "'enhancement', by contrast (to therapy), is the directed use of biotechnological power to alter, by direct intervention, not disease processes but the 'normal' workings of the human body and psyche, to augment or improve their native capacities and performances" - only to question this definition later on and dismiss the whole concept of enhancement as being "finally inadequate to the moral analysis" (PCB 2003). Indeed, making a distinction between "therapy" and "enhancement" may be problematic as long as one does not limit the latter term to "non-therapeutic enhancement", i.e. enhancement of healthy persons. For this chapter, we will stick with this distinction for the sake of simplicity, and add some clarification on its uses and interpretation whenever necessary. We will therefore go along with the terminology laid down by Eric Juengst and Erik Parens in 1998 and apply the term enhancement to characterise "interventions designed to
14 Philosophy Unit of the Royal Institute of Technology, Stockholm, Sweden.
improve human form or functioning beyond what is necessary to sustain or restore good health" (Parens 1995a; Juengst 1998). This concept will also require us to lead the ethical discussion of the concept of enhancement on at least two different levels: the individual level (including the ethics of self-improvement), and the societal level (including the goals of medicine and the goals of society).
As opposed to psychopharmacology (see Chapter 1), there is no method of enhancement by neural implants that is already in use, although technologically, as one can see from the recent developments in the field described above, it would be possible today: Artificial limbs can be controlled even with non-invasive Human Machine Interfaces "by thought alone"; non-verbal, invisible communication can be achieved with Human Computer Interfaces; and by electrical stimulation of specific areas in the brain, mood can be altered and confidence be boosted. One of the reasons why such direct brain-computer interfaces are not yet being considered in normal humans is that these technologies still require a more or less invasive approach and relatively coarse arrays of electrodes implanted inside the human central nervous system.
With the improvement of telemetric stimulation, avoiding the necessity of a direct electrode-tissue contact, sensory enhancement may be not so farfetched an idea anymore. While some sensory enhancements do not have to involve invasive approaches at all (like ultrasound hearing or infrared vision) and will only be accepted in form of implants if they are absolutely safe, any effective cognitive or memory enhancement may be a different story both in terms of its technical establishment and of the acceptance of involved risks.
An important issue will be the reversibility of such an enhancement. Electrical stimulation can be interrupted at any given time by the implanted person him- or herself, provided that the stimulation has not led to side effects that make it impossible for the carrier of the implant to use the appropriate external switch.
Another issue that will have its part in the discussion is intention. With every new step in the development of technologies, there will be a potential for abuse. To many, enhancement by connecting electronic devices to the human body and brain will appear, at least prima facie, morally suspect. Others have argued that these technologies offer an opportunity to make life even more worth living by getting smarter, being happier, living longer -provided that society would respond democratically to the implicit social challenges (Hughes 2004).
Was this article helpful?
Have you ever been envious of people who seem to have no end of clever ideas, who are able to think quickly in any situation, or who seem to have flawless memories? Could it be that they're just born smarter or quicker than the rest of us? Or are there some secrets that they might know that we don't?